5Setting Comfort Levels and Loudness Balancing

FThe upper level: C, M, and MCL

The upper-stimulation level caps the electrical dynamic range; it is the M level (most comfortable) for Advanced Bionics, the C level for Nucleus, and the maximum comfort level for MED-EL. Provision of optimal upper levels is one of the most important parts of programming: underestimating them sacrifices speech recognition, sound quality, and self-voice monitoring, while overestimating them causes discomfort, can degrade speech recognition, and can foster an aversive reaction to the implant. The loudness target the recipient is taken to differs by maker: most comfortable for Advanced Bionics, loud but comfortable for Cochlear, and loud but not uncomfortable for MED-EL. The maximum level a recipient will tolerate typically rises over the first weeks to months of use, and there is no universal set of levels, so upper levels must be individualised and revisited. Real-world loudness is driven by summation across many simultaneously stimulated channels, so frequency-specific loudness judged channel-by-channel may not predict the optimal everyday upper level.[2020][2002]

CLoudness scaling and global adjustment

In psychophysical loudness scaling the recipient points to a loudness chart, for example a 0 to 10 scale from off and just noticeable up through most comfortable to too loud, as the audiologist raises the stimulus to the maker's recommended upper-level descriptor. Channel-by-channel scaling is time consuming and may not reflect real-world summed loudness, so a global-adjustment method is widely used: T levels are temporarily dropped below audibility, the upper levels are superimposed, and the audiologist raises overall volume in live-speech mode to a comfortable level. Recipients often declare the level optimal too early, so after the first comfortable report the audiologist adds a small further increase to test whether clarity improves, then backs off if quality or comfort deteriorates. Once the upper levels are set globally, T levels are returned to their measured values, with a small global T decrease if a static or buzz or ambient-noise complaint appears. The slope of loudness growth (the Nucleus Q value, default 20, adjustable 10 to 50) controls how much of the electrical dynamic range is given to the top of the input range; lowering it steepens growth and adds salience for soft inputs, but it rarely needs changing when levels are set well.[2020][2002]

CSweeping and loudness balancing across the array

Sweeping at the upper level (beeping each channel in turn) confirms that no electrode is uncomfortably loud, that pitch rises smoothly apex-to-base, and that no channel produces poor sound quality or non-auditory side effects. Loudness balancing presents a stimulus successively to two or more channels at a fixed percentage of the dynamic range to confirm they are equally loud; equal-loudness percepts across the array are tied to better sound quality and speech recognition. A common method balances two electrodes at a time starting apically, for example 22 versus 21: if 21 is louder, its level is reduced, the pair is rechecked, then 21 is balanced against 20, and so on across the array. The adjustment is always made to the second electrode of each pair, because the first has already been balanced to the electrodes preceding it. If channels are not balanced, the loudest channel dominates the overall percept; the user then lowers global volume so that channel is tolerable, starving the remaining channels of adequate stimulation and erasing the natural intensity differences between phonemes. Sweeping and balancing can be done across all channels or just two to four at a time when finer responses are needed.[2020][2002]

TShaping the profile and the danger of levels set too high

The upper-level profile across the array is normally flatter than the T-level profile, so after a global increase the audiologist smooths the upper levels rather than leaving abrupt channel-to-channel jumps, and never simply flattens them all to one value. Where basal T levels are high, basal upper levels are also higher, but a slight basal taper makes the upper profile flatter than the T profile, leaving a narrower dynamic range basally than apically. Tilting (tapering low- or high-frequency upper levels) addresses sound-quality complaints, for example reducing high-frequency upper levels when sounds are too sharp or low-frequency levels when there is too much bass. Levels set too high cause discomfort and an aversive reaction, can degrade rather than improve speech recognition, and waste the electrical dynamic range. Objective measures (electrically evoked stapedius reflex thresholds and ECAP) help bracket the upper level, particularly in young children who cannot reliably scale loudness, guarding against overstimulation. An overly loud channel can also drive excessive current that raises channel interaction, reinforcing the value of careful balancing.[2020][2014]